perf: add ability to sample interrupted machine state

Enable capture of interrupted machine state in each sample.

Registers to sample are passed per event in the sample_regs_intr bitmask.

To sample interrupt machine state, the PERF_SAMPLE_INTR_REGS must be passed in sample_type.

The list of available registers is arch dependent and provided by asm/perf_regs.h

Registers are laid out as u64 in the order of the bit order of sample_intr_regs.

Signed-off-by Stephane Eranian <[email protected]>

---
 include/linux/perf_event.h      |  7 +++++--
 include/uapi/linux/perf_event.h | 14 ++++++++++++-
 kernel/events/core.c            | 44 +++++++++++++++++++++++++++++++++++++++--
 3 files changed, 60 insertions(+), 5 deletions(-)

include/linux/perf_event.h +3 -2

index 707617a..4970c1d 100644
@@ -78,7 +78,7 @@ struct perf_branch_stack {
 	struct perf_branch_entry	entries[0];
 };
 
-struct perf_regs_user {
+struct perf_regs {
 	__u64		abi;
 	struct pt_regs	*regs;
 };

@@ -595,7 +595,8 @@ struct perf_sample_data {
 	struct perf_callchain_entry	*callchain;
 	struct perf_raw_record		*raw;
 	struct perf_branch_stack	*br_stack;
-	struct perf_regs_user		regs_user;
+	struct perf_regs		regs_user;
+	struct perf_regs		regs_intr;
 	u64				stack_user_size;
 	u64				weight;
 	/*

Enable capture of interrupted machine state in each sample.

Enable capture of interrupted machine state in each sample.

Registers to sample are passed per event in the sample_regs_intr bitmask.

To sample interrupt machine state, the PERF_SAMPLE_INTR_REGS must be passed in sample_type.

The list of available registers is arch dependent and provided by asm/perf_regs.h

Registers are laid out as u64 in the order of the bit order of sample_intr_regs.

Enable capture of interrupted machine state in each sample.

Registers to sample are passed per event in the sample_regs_intr bitmask.

To sample interrupt machine state, the PERF_SAMPLE_INTR_REGS must be passed in sample_type.

The list of available registers is arch dependent and provided by asm/perf_regs.h

Registers are laid out as u64 in the order of the bit order of sample_intr_regs.

Enable capture of interrupted machine state in each sample.

Registers to sample are passed per event in the
sample_regs_intr bitmask.

To sample interrupt machine state, the
PERF_SAMPLE_INTR_REGS must be passed in
sample_type.

index 57e9190..018939c 100644

kernel/events/core.c +1 -1

@@ -4296,7 +4296,7 @@ perf_output_sample_regs(struct perf_output_handle *handle,
 	}
 }
 
-static void perf_sample_regs_user(struct perf_regs_user *regs_user,
+static void perf_sample_regs_user(struct perf_regs *regs_user,
 				  struct pt_regs *regs)
 {
 	if (!user_mode(regs)) {

PEBS can capture machine state regs at retirement of the sampled instructions. When precise sampling is enabled on an event, PEBS is used, so substitute the interrupted state with the PEBS state. Note that not all registers are captured by PEBS. Those missing are replaced by the interrupt state counter-parts.

Signed-off-by Stephane Eranian <[email protected]>

---
 arch/x86/kernel/cpu/perf_event_intel_ds.c | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

arch/x86/kernel/cpu/perf_event_intel_ds.c +7 -0

index 980970c..d612bcd 100644
@@ -925,6 +925,25 @@ static void __intel_pmu_pebs_event(struct perf_event *event,
 	regs.bp = pebs->bp;
 	regs.sp = pebs->sp;
 
+	if (sample_type & PERF_SAMPLE_REGS_INTR) {
+		regs.ax = pebs->ax;
+		regs.bx = pebs->bx;
+		regs.cx = pebs->cx;
+		regs.si = pebs->si;
+		regs.di = pebs->di;
+

Add the infrastructure to setup, collect and report the interrupt machine state regs which can be captured by the kernel.

Signed-off-by Stephane Eranian <[email protected]>

---

tools/perf/util/session.c +12 -0

+static const char *regs_abi[]= {
+ [PERF_SAMPLE_REGS_ABI_NONE] = "none",
+ [PERF_SAMPLE_REGS_ABI_32] = "32-bit",
+ [PERF_SAMPLE_REGS_ABI_64] = "64-bit",
+};
+
+static inline const char *regs_dump_abi(struct regs_dump *d)
+{
+	if (d->abi > PERF_SAMPLE_REGS_ABI_64)
+		return "???";
+	return regs_abi[d->abi];
+}

Add the infrastructure to setup, collect and report the interrupt machine state regs which can be captured by the kernel.

index 8606175..00e45d1 100644

tools/perf/util/evsel.c +6 -0

@@ -626,6 +626,11 @@ void perf_evsel__config(struct perf_evsel *evsel, struct record_opts *opts)
 	if (opts->call_graph_enabled)
 		perf_evsel__config_callgraph(evsel, opts);
 
+	if (opts->sample_intr_regs) {
+		attr->sample_regs_intr = PERF_REGS_MASK;
+		perf_evsel__set_sample_bit(evsel, REGS_INTR);
+	}
+
 	if (target__has_cpu(&opts->target))
 		perf_evsel__set_sample_bit(evsel, CPU);
 

@@ -979,6 +984,7 @@ static size_t perf_event_attr__fprintf(struct perf_event_attr *attr, FILE *fp)
 	ret += PRINT_ATTR_X64(branch_sample_type);
 	ret += PRINT_ATTR_X64(sample_regs_user);
 	ret += PRINT_ATTR_U32(sample_stack_user);
+	ret += PRINT_ATTR_X64(sample_regs_intr);
 
 	ret += fprintf(fp, "%.60s\n", graph_dotted_line);
 

Quoted 3 lines · Stephane Eranian

Add the infrastructure to setup, collect and report the interrupt machine state regs which can be captured by the kernel.

SNIP

Quoted 23 lines

index 8606175..00e45d1 100644

tools/perf/util/evsel.c +6 -0

@@ -626,6 +626,11 @@ void perf_evsel__config(struct perf_evsel *evsel, struct record_opts *opts)
      if (opts->call_graph_enabled)
              perf_evsel__config_callgraph(evsel, opts);
 
+     if (opts->sample_intr_regs) {
+             attr->sample_regs_intr = PERF_REGS_MASK;
+             perf_evsel__set_sample_bit(evsel, REGS_INTR);
+     }
+
      if (target__has_cpu(&opts->target))
              perf_evsel__set_sample_bit(evsel, CPU);
 

@@ -979,6 +984,7 @@ static size_t perf_event_attr__fprintf(struct perf_event_attr *attr, FILE *fp)
      ret += PRINT_ATTR_X64(branch_sample_type);
      ret += PRINT_ATTR_X64(sample_regs_user);
      ret += PRINT_ATTR_U32(sample_stack_user);
+     ret += PRINT_ATTR_X64(sample_regs_intr);
 
      ret += fprintf(fp, "%.60s\n", graph_dotted_line);
 

tools/perf/util/evsel.c int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event, · line 1477

Quoted 5 lines

tools/perf/util/evsel.c +1 -0

@@ -1468,6 +1474,23 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
              array++;
      }
 
+     data->intr_regs.abi = PERF_SAMPLE_REGS_ABI_NONE;

not necessary, data is zeroed in the top

Add the infrastructure to setup, collect and report the interrupt machine state regs which can be captured by the kernel.

Signed-off-by Stephane Eranian <[email protected]>

---
 tools/perf/perf.h         |  1 +
 tools/perf/util/event.h   |  7 +++++++
 tools/perf/util/evsel.c   | 25 ++++++++++++++++++++++++-
 tools/perf/util/session.c | 26 ++++++++++++++++++++++++++
 4 files changed, 58 insertions(+), 1 deletion(-)

tools/perf/perf.h +0 -0

index 510c65f..309d956 100644

Add -I option to capture machine state registers at interrupt.

Use perf report -D to display the sampled register values in each sample.

Signed-off-by Stephane Eranian <[email protected]>

---
 tools/perf/builtin-record.c | 2 ++
 1 file changed, 2 insertions(+)

tools/perf/builtin-record.c +2 -0

index 378b85b..17c45f8 100644
@@ -876,6 +876,8 @@ const struct option record_options[] = {
 		    "sample transaction flags (special events only)"),
 	OPT_BOOLEAN(0, "per-thread", &record.opts.target.per_thread,
 		    "use per-thread mmaps"),
+	OPT_BOOLEAN('I', "sample-intr-regs", &record.opts.sample_intr_regs,
+		    "Sample machine register on interrupt"),
 	OPT_END()
 };

Hi Stephane,

Quoted 1 line · Stephane Eranian

Add -I option to capture machine state registers at interrupt.

You need to add it in the Documentation too.

Thanks,
Namhyung

Quoted 21 lines

Use perf report -D to display the sampled register values in each sample.

Signed-off-by Stephane Eranian <[email protected]>

---
 tools/perf/builtin-record.c | 2 ++
 1 file changed, 2 insertions(+)

tools/perf/builtin-record.c +2 -0

index 378b85b..17c45f8 100644
@@ -876,6 +876,8 @@ const struct option record_options[] = {
                  "sample transaction flags (special events only)"),
      OPT_BOOLEAN(0, "per-thread", &record.opts.target.per_thread,
                  "use per-thread mmaps"),
+     OPT_BOOLEAN('I', "sample-intr-regs", &record.opts.sample_intr_regs,
+                 "Sample machine register on interrupt"),
      OPT_END()
 };

This short patch series add the ability to sample the interrupted machine state for each hardware sample. This is useful to analyze the state after certain events, for instance for function value profiling after a call instruction.

The patch extends the interface with a new PERF_SAMPLE_REGS_INTR smaple_type flag. The register to sample can be named in the sample_regs_intr bitmask for each event. The name and bit position for each register is architecture dependent and provided, just like for PERF_SAMPLE_REGS_USER by asm/perf_regs.h.

The support is similar to PERF_SAMPLE_REGS_USER.

On Intel x86, the series includes support for capturing the PEBS state as well. When precise sampling is used, the interrupted state is collect from the PEBS records, at least partially. The PEBS machine state is a subset of the machine state.

The series provides access to this new feature in perf record with the -I option. It is possible to display the sampled register values using perf report -D.

This patch series is the fundation for a future series adding function value profiling.

Stephane Eranian (4):
perf: add ability to sample machine state on interrupt perf/x86: add support for sampling PEBS machine state registers perf tools: add core support for sampling intr machine state regs perf record: add new -I option to sample interrupted machine state

 arch/x86/kernel/cpu/perf_event_intel_ds.c |   19 +++++++++++++
 include/linux/perf_event.h                |    7 +++--
 include/uapi/linux/perf_event.h           |   14 ++++++++-
 kernel/events/core.c                      |   44 +++++++++++++++++++++++++++--
 tools/perf/builtin-record.c               |    2 ++
 tools/perf/perf.h                         |    1 +
 tools/perf/util/event.h                   |    7 +++++
 tools/perf/util/evsel.c                   |   23 ++++++++++++++-
 tools/perf/util/session.c                 |   26 +++++++++++++++++
 9 files changed, 137 insertions(+), 6 deletions(-)

Quoted 2 lines

This patch series is the fundation for a future series adding function value profiling.

Probably a silly question: what is the function value profiling?

Em Wed, Jul 09, 2014 at 05:26:20PM +0900, Namhyung Kim escreveu:

Quoted 5 lines · Stephane Eranian

Quoted 2 lines

This patch series is the fundation for a future series adding function value profiling.

Probably a silly question: what is the function value profiling?

I think he wants to sample the values of parameters and local variables, etc, right?

http://lists.cs.uiuc.edu/pipermail/llvmdev/2013-June/062950.html lists some things you can do with that.

Stephane, is that what you mean by "value profiling"?

Value profiling means sampling the values of function arguments passed in registers.

Quoted 2 lines

Value profiling means sampling the values of function arguments passed in registers.

You can use it for more, e.g. it's also useful to sample loop iterations.

-Andi

--
[email protected] -- Speaking for myself only